Summary of Skeleton: a New Framework For Accelerating Language Models Via Task Neuron Localized Prompt Tuning, by Nakyeong Yang et al.
Skeleton: A New Framework for Accelerating Language Models via Task Neuron Localized Prompt Tuning
by Nakyeong Yang, Jiwon Moon, Junseok Kim, Yunah Jang, Kyomin Jung
First submitted to arxiv on: 18 Apr 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel prompt tuning framework called Skeleton that efficiently utilizes a language model by retaining only task-relevant neurons using an explainability method. The framework enables solving various tasks with a single language model while accelerating inference speed during application procedures. By prepending adequate task-specific prompt tokens, the method achieves comparable performance to existing prompt tuning methods while reducing memory and time complexity. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making computers faster at doing certain tasks. Right now, when we teach them new things, they use a lot of brain power (memory) and take a long time to figure out what to do. The researchers developed a way to make the computer use only the parts of its “brain” that are actually needed for each task. This makes it work faster and more efficiently, without sacrificing performance. |
Keywords
» Artificial intelligence » Inference » Language model » Prompt
